Patient Digital Twin — the time is NOW!
Every patient already has her health information in a tattered paper folder in her doctor’s office — this IS a “paper twin” of the patient; it contains quarterly clinic visit notes, blood work results, meds, etc. Most health systems have digitized these paper records placing it in the cloud for anytime access of the patient and the physician. This is a form of patient “digital twin”. You see, a CGI generated “avatar” of the patient with data displayed on the avatar is not necessary for it to be a digital twin — this is an unnecessary hang-over induced by Industrial digital twins where pictures of large turbines color coded with the current pressure, temperature, etc. are shown in every digital twin dashboard!
Digital twin is not only about a repository of patient and related data — true, this patient digital twin will receive data and update itself in “real-time” (may be every quarter with the clinic visit) but a digital twin can and should do more.
Say, the patient catches Covid and is admitted into her local hospital. When being in the hospital ward, vital signs such as heart rate, oxygenation, blood pressure, etc. are monitored using “hospital IoT”. In most cases, the data is for healthcare professionals to see in tiny local displays; in others, it is connected to hospital IT and the patient digital twin in the cloud is updated with these data in “real-time” (here, may be every minute).
Now the patient digital twin can be made “smart” with simple machine learning (ML) that takes the vital signs together and predicts severity and duration before emergency. Unfortunately, in this patient’s case, ICU admission is the order and while in the ICU, additional vital signs are acquired more frequently and more advanced ML (such as CAUSAL machine learning) algorithms are applied to assist the physician to determine the root-causes of the patient’s worsening condition. Physician can also use the ML outputs (severity and duration) to help triage ICU candidates in the case of the next pandemic.
Fortunately, our patient gets well and is discharged to her home in her small town; she lives alone with family members far away. Her grandson in the big city hundreds of miles away is anxious about grandma’s well-being and perhaps feels a little guilty for abandoning her to fend for herself.
Grandma’s digital twin is now developed further with IoT in her home; video cameras and well-developed video analytics software computes the number of steps she walks per day, how may time she cooked and went to the bathroom and social interactions from voice prints extracted by an Alexa-type device in her home. (If grandma finds cameras spying on her as unacceptable, most of her activities of daily living can be deduced from appropriate IoT sensors).
Such a grandma digital twin will be a great boon to the loving grandson (who is now willing to pay a monthly subscription for this service!) as well as her family doctor whose healthcare associate (may be at a remote location) can monitor weekly aggregates, process alarms, send out updates, engage grandma once in a while and react to worsening medical condition locally and quickly by bringing the family doctor into the loop.
A recent article from PubMed (“Digital Twins: From Personalised Medicine to Precision Public Health” (J Pers Med. 2021 Aug) had this to say:
“Applied to medicine and public health, digital twin technology can drive a much-needed radical transformation of traditional electronic health/medical records (focusing on individuals) and their aggregates (covering populations) to make them ready for a new era of precision (and accuracy) medicine and public health”.
All the pieces are in place! What we need NOW is for some major corporations like Walmart Healthcare or Amazon Health or major hospital systems to step up, assemble the pieces and deploy patient digital twins widely . . .
Dr. PG Madhavan
https://www.linkedin.com/in/pgmad/
#digitaltwin #patientdigitaltwin #ML #CausalML #IoT
